Running head: Lianas always outperform trees (Pasquini et al.) 1 Lianas always outperform tree seedlings regardless of soil nutrients: results from a long-1 term fertilization experiment 2 3 Sarah C. Pasquini1,3, S. Joseph Wright2 and Louis S. Santiago1,2 4 5 1Department of Botany and Plant Sciences, University of California, 2150 Batchelor Hall, 6 Riverside, CA 92521, USA 7 8 2Smithsonian Tropical Research Institute, Apartado 0843-03092, Balboa, Republic of Panamá 9 10 3Corresponding author, email: sarah.pasquini@email.ucr.edu11 Running head: Lianas always outperform trees (Pasquini et al.) 2 Abstract 12 Lianas are a prominent growth form in tropical forests and there is compelling evidence 13 that they are increasing in abundance throughout the Neotropics. While recent evidence shows 14 that soil resources limit tree growth even in deep shade, the degree to which soil resources limit 15 lianas in forest understories, where they coexist with trees for decades, remains unknown. 16 Regardless, the physiological underpinnings of soil resource limitation in deeply shaded tropical 17 habitats remain largely unexplored for either trees or lianas. Theory predicts that lianas should 18 be more limited by soil resources than trees because they occupy the quick-return end of the 19 “leaf economic spectrum” characterized by high rates of photosynthesis, high specific leaf area, 20 short leaf life span, affinity to high-nutrient sites, and greater foliar nutrient concentrations. To 21 address these issues, we asked whether soil resources (nitrogen, phosphorus, and potassium), 22 alone or in combination, applied experimentally for more than a decade would cause significant 23 changes in morphology or physiology of tree and liana seedlings in a lowland tropical forest. 24 We found evidence for the first time that phosphorus limits the photosynthetic performance of 25 both trees and lianas in deeply shaded understory habitats. More importantly, lianas always 26 showed significantly greater photosynthetic capacity, quenching, and saturating light levels 27 compared to trees across all treatments. We found little evidence for nutrient × growth form 28 interactions, indicating that lianas were not disproportionately favored in nutrient-rich habitats. 29 Tree and liana seedlings differed markedly for six key morphological traits demonstrating that 30 architectural differences occurred very early in ontogeny prior to lianas finding a trellis (all 31 seedlings were self-supporting). Overall, our results do not support nutrient loading as a 32 mechanism of increasing liana abundance in the Neotropics. Rather, our finding that lianas 33 always outperform trees, in terms of photosynthetic processes and under contrasting rates of 34 Running head: Lianas always outperform trees (Pasquini et al.) 3 resource supply of macronutrients, will allow lianas to increase in abundance if disturbance and 35 tree turnover rates are increasing in Neotropical forests as has been suggested. 36 37 Keywords: Barro Colorado Nature Monument, Panama; Chlorophyll Fluorescence; Fertilization; 38 Neotropics; Nitrogen; Nutrient Limitation; Phosphorus; Photosynthetic Performance; Plant 39 Architecture; Plant Morphology; Potassium; Tropical Forest 40 41 Introduction 42 Lianas and trees are the two dominant plant growth forms in tropical forests and there is a 43 growing body of evidence suggesting that lianas are increasing relative to trees in Neotropical 44 forests (Phillips et al. 2002, Benítez-Malvido and Martínez-Ramos 2003, Wright, S. J. et al. 45 2004, Chave et al. 2008, Foster et al. 2008, Schnitzer and Bongers 2011, Yorke et al. 2013, 46 Schnitzer 2015). We are not sure why. Regardless, these increases in liana abundance will 47 almost certainly have important consequences for forest biodiversity and global carbon budgets 48 (Bunker et al. 2005, Schnitzer and Carson 2010, Schnitzer and Bongers 2011, Schnitzer et al. 49 2014). Indeed, Schnitzer and Carson (2010) and Schnitzer et al. (2014) demonstrated 50 unequivocally that when lianas increase in abundance and displace trees, forest-wide above-51 ground carbon storage can be reduced by as much as 18%. While exceptions exist, lianas are 52 typically a fast-growing, light-limited growth form associated with high-light, nutrient-rich, and 53 disturbed habitats, including forest edges, canopy gaps, and logged forests (Putz 1984, Schnitzer 54 et al. 2000, Schnitzer and Carson 2010). Moreover, they can represent more than a third of all 55 woody species in tropical forests (Pérez-Salicrup et al. 2001, Gentry 2009, Schnitzer et al. 2012). 56 Lianas differ from trees in critical patterns of biomass allocation and other key life 57 history traits. For example, in their climbing form, lianas rely on other vegetation as trellises to 58 Running head: Lianas always outperform trees (Pasquini et al.) 4 gain access to the canopy, and thus they typically invest proportionally less resources into woody 59 stem tissue than trees, and proportionally more resources toward leaves and roots (Putz 1983, 60 Suzuki 1987, Castellanos et al. 1989, Niklas 1994, Gerwing and Farias 2000, Hättenschwiler 61 2002, Santiago and Wright 2007). This biomass allocation pattern of canopy-level lianas (lianas 62 with at least partial foliage in the forest canopy) results in lianas having greater specific leaf area 63 (SLA) and photosynthetic rates (Amax) than trees (Zhu and Cao 2009, Han et al. 2010, Zhu and 64 Cao 2010, Asner and Martin 2012, Santiago et al. 2015). The abundance of canopy-level lianas 65 often increases with soil fertility (e.g., Proctor et al. 1983, Putz 1983, 1985, Putz and Chai 1987, 66 Balfour and Bond 1993, Bruijnzeel and Proctor 1995) and lianas show higher foliar nutrient 67 concentrations compared to trees (Cai and Bongers 2007, Kusumoto and Enoki 2008, Zhu and 68 Cao 2010, Asner and Martin 2012). This suggests that canopy-level lianas are far more nutrient-69 limited than trees, yet the few in situ experimental nutrient enrichment studies available have 70 found either modest support for this (Hättenschwiler 2002) or no differences at all between the 71 life forms (Cai et al. 2008). 72 While canopy-level trees and lianas differ strongly in terms of leaf traits, morphology, 73 and physiology, the degree to which seedlings differ is unclear. In early ontogenetic stages, 74 lianas typically exist without a trellis and both trees and lianas have to survive for years within 75 deeply shaded understory habitats; under these conditions they appear strikingly similar in terms 76 of morphology and architecture (Putz 1983). Thus, there appears to be broad overlap in the 77 patterns of growth, survival, and habitat preferences of the seedlings of both lianas and trees 78 (Gilbert et al. 2006). Consequently, contrasting resource uptake and allocation may not occur in 79 early developmental stages where light remains the primary limiting resource, and most 80 differences between trees and lianas may only develop late in ontogeny. 81 Running head: Lianas always outperform trees (Pasquini et al.) 5 While light may be the most limiting resource in tropical forest understory habitats, it has 82 recently become clear that seedlings of some woody species are also co-limited by soil nutrients. 83 Limitation varies among species (Denslow et al. 1987) and among soil resources including 84 nitrogen (N), phosphorus (P), or potassium (K) and in some cases limitation is caused 85 simultaneously by multiple soil resources (Bloom et al. 1985, Ceccon et al. 2004, Holste et al. 86 2011, Wright et al. 2011, Pasquini and Santiago 2012, Santiago et al. 2012). Indeed, recent work 87 on the tree seedling, Alseis blackiana (Helms.; Rubiaceae), demonstrated that photosynthesis, 88 stomatal conductance, and photosynthetic yield were limited by N, P, and K, respectively, even 89 in deep shade (Pasquini and Santiago 2012). In a separate study at the same site K limited tree 90 seedling growth (Santiago et al. 2012). For liana seedlings, however, both the degree of nutrient 91 limitation, as well as whether liana seedlings are more or less limited by soil resources than tree 92 seedlings, remains unknown. 93 Here we test the hypothesis in situ that nutrients limit photosynthetic physiology of liana 94 seedlings to a greater degree than tree seedlings. If lianas and trees differ, we would demonstrate 95 that in spite of their apparent similarities in seedling morphology, physiological divergence 96 happens early in ontogeny, and if not, then physiological differences must develop after they find 97 a trellis and begin ascending into the canopy. Furthermore, testing our hypothesis may provide 98 insight into the underlying mechanism for the increase in lianas in many Neotropical forests. 99 Nutrient deposition, particularly of nitrogen, is increasing throughout the tropics (reviewed by 100 Hedin et al. 2009, Hietz et al. 2011), tree turnover rates also appear to be increasing (Phillips et 101 al. 2004) as are rates of human disturbances and deforestation (e.g., reviewed by Laurance 2008, 102 Wright 2010). All of these are likely to favor lianas particularly if they gain an advantage early 103 in ontogeny. 104 Running head: Lianas always outperform trees (Pasquini et al.) 6 To test our hypothesis we are using a fully factorial experiment where N, P, and K have 105 been added to large replicated forest plots for more than a decade. We compare how soil 106 resources impact the physiology and morphology of seedlings of a phylogenetically diverse 107 group of lianas and trees from 13 plant families. We hypothesize that: 1) Lianas will show 108 greater responses to soil nutrients than trees because of their ability to allocate more to growth 109 versus structural support, 2) Lianas will be limited by different soil resources than trees, and 3) 110 Lianas in very early developmental stages prior to acquiring a trellis will have contrasting 111 patterns of plant architecture (e.g., internode length and leaf angle) compared to trees. Our goal 112 is to determine whether liana and tree seedlings are constrained by the same or different 113 resources or combinations of resources and link this to key aspects of photosynthetic physiology 114 and seedling architecture. Ultimately we link our findings back to recent evidence that strongly 115 suggests lianas are not only increasing in abundance throughout the Neotropics but also altering 116 patterns of carbon storage and sequestration (e.g., Schnitzer et al. 2014, Schnitzer 2015). 117 118 Materials and Methods 119 Study site 120 We performed this research in seasonally moist, semi-deciduous, tropical forest located 121 on the Gigante Peninsula (9°06’31” N, 79°50’37” W) within the Barro Colorado Nature 122 Monument (BCNM) in central Panama (Appendix A: Fig. A1). The dry season occurs between 123 January and April during which less than 10% of the 2600 mm of average annual rainfall occurs. 124 Our investigation took place from March through April 2010. Soils on the Gigante Peninsula are 125 Oxisols and Inseptisols similar to Typic Eutrudox soils on adjacent Barro Colorado Island 126 (Turner et al. 2012, B. L. Turner personal communication). In terms of N, P, and K availability, 127 Running head: Lianas always outperform trees (Pasquini et al.) 7 soils at this site are relatively fertile for lowland tropical soils (Yavitt et al. 2009, Wright et al. 128 2011). Tree composition and stature (tree heights up to 45 m) in this forest are characteristic of 129 mature (> 200 y) tropical secondary forest in central Panama (Wright et al. 2011). 130 Experimental design 131 We used a long-term nutrient addition experiment where N, P, and K have been added in 132 a full 2 × 2 × 2 factorial design with 4 replicates of each of 8 treatments (control, N, P, K, NP, 133 NK, PK, NPK). The 4 replicates were placed perpendicular to a slight topographical gradient 134 (36 m in elevation from southwest to northeast corner of site) because tree distributions and soil 135 properties parallel this gradient (Yavitt et al. 2009, Wright et al. 2011). We used a balanced, 136 incomplete-block design, where N, P, K, and NPK treatments were blocked versus NP, NK, PK, 137 and control treatments within each replicate (Wright et al. 2011, Pasquini and Santiago 2012, 138 Santiago et al. 2012). This design minimizes uncontrolled error due to spatial heterogeneity and 139 allows evaluation of main effects and two-way interactions, but limits power to evaluate the 140 three-way interaction (Winer et al. 1991). Nutrients were added 4 times annually during the wet 141 season for a total of 125 kg N ha-1 y-1 as coated urea [(NH2)2CO], 50 kg P ha-1 y-1 as triple super 142 phosphate [Ca(H2PO4)2•H2O] and 50 kg K ha-1 y-1 as potassium chloride (KCl) starting in 1998 143 (12 years of nutrient addition). The 32 experimental plots were each 40 × 40 m in area and were 144 separated by at least 40 m to minimize nutrient leaching into neighboring plots, with the 145 exception of two plots separated by 20 m and located on opposite sides of a 3-m deep stream. In 146 this same study site, long-term N fertilization lead to increased soil acidity (0.8 unit decrease in 147 soil pH; Corre et al. 2010), which may affect availability of P and other soil nutrients. 148 Species 149 Running head: Lianas always outperform trees (Pasquini et al.) 8 We selected seven liana and six tree species from 13 plant families because they were 150 common in the study plots. The seven lianas were Bauhinia guianenses Aubl. (Fabaceae – 151 Caesalpinioideae), Coccoloba parimensis Benth. (Polygonaceae), Doliocarpus dentatus (Aubl.) 152 Standl. (Dilleniaceae), Maripa panamensis Hemsl. (Convolvulaceae), Paullinia fibrigera Radlk. 153 (Sapindaceae), Phryganocydia corymbosa (Vent.) Bureau ex. K. Schum (Bignoniaceae) and 154 Prionostemma aspera (Lam.) Miers. (Celastraceae). The trees were Alseis blackiana Hemsl. 155 (Rubiaceae), Desmopsis panamensis (B. L. Rob.) Saff. (Annonaceae), Heisteria concinna Standl. 156 (Olacaceae), Oenocarpus mapora H. Karst. (Arecaceae), Sorocea affinis Hemls. (Moraceae) and, 157 Tetragastris panamensis (Engler) Kuntze (Burseraceae). Nomenclature follows Garwood 158 (2009). Individual seedlings were chosen haphazardly based on the first sightings of the study 159 species within each plot. All liana seedlings were self-supporting (free-standing) and did not 160 exhibit searcher shoots (sensu Putz and Holbrook 2009). 161 Physiological measurements 162 Chlorophyll fluorescence measurements were used because they are highly correlated 163 with carbon assimilation rates (especially maximum electron transport, Maxwell and Johnson 164 2000) and we confirmed this relationship for one of our focal species (Alseis blackiana; Pasquini 165 and Santiago 2012). We measured chlorophyll fluorescence of mature, fully expanded leaves 166 using a photosynthesis yield analyzer (Mini-PAM, Heinz Walz GmbH, Effeltrich, Germany). 167 We sampled one leaf from one individual of the 13 species in each of the 32 plots (mean leaves 168 sampled per plot = 12.4, total leaves sampled = 397). We constructed chlorophyll fluorescence 169 light response curves using photon flux density (PFD) values of 0, 34, 97, 202, 324, 499, 700, 170 1067 and 1471 μmol m-2 s-1 to slowly bring the light level up to the highest light level. We 171 measured the electron transport rate (ETR), which is an in vivo measure of overall capacity to 172 Running head: Lianas always outperform trees (Pasquini et al.) 9 provide energy to photosynthetic carboxylation reactions. We also measured photochemical 173 quenching (qp). qp is the proportion of open photosystem II (PSII) reaction centers and is a proxy 174 of the efficiency of PSII. We obtained the maximum electron transport rate (ETRmax) and P at 175 the highest light level (PFD = 1471 μmol m-2 s-1). ETR was determined as: 176 177 (1) 178 179 where ܨ୫ᇱ is maximal fluorescence measured by a saturation pulse at each light level and ܨୱ is 180 steady-state fluorescence. f is a factor that represents the partitioning of photons between 181 photosystems II and I (PSII and PSI) and is assumed to be 0.5, which indicates equal distribution 182 of excitation energy between the two photosystems (Maxwell and Johnson 2000), and α 183 represents the fraction of photons absorbed by a leaf and is assumed to be 0.84 as an average for 184 a variety of C3 leaves (Björkman and Demmig 1987, Stemke and Santiago 2011). 185 Photochemical quenching was determined as: 186 187 (2) 188 189 where ܨ଴ᇱ is minimum fluorescence of each illuminated sample determined during a brief dark 190 interval following a saturation pulse (see Fig. 1 for an example of a fluorescence light response 191 curve). 192 Morphological measurements 193 We measured leaf angle, leaf thickness, internode length, and petiole length to 194 characterize seedling morphology. Crown depth, perpendicular crown widths, and seedling 195 Running head: Lianas always outperform trees (Pasquini et al.) 10 height were measured and used to calculate crown depth and crown area, relative to height. Leaf 196 angle was measured using a protractor with a weighted thread as the angle of the leaf measured 197 along the midvein from petiole attachment to leaf tip where a 90° leaf angle is parallel to the 198 ground and perpendicular to the main stem (leaf angle > 90° indicates that leaf at an obtuse angle 199 relative to the ground). We measured leaf thickness on an area of the leaf without major veins 200 using a digital micrometer (IP 65, Mitutoyo Corp., Mizonokuchi, Japan). Additionally, one leaf 201 from each seedling was collected and measured for leaf area (leaf petiole was removed) using a 202 leaf area meter (LI-3100, Li-Cor Biosciences, Inc., Lincoln, Nebraska, USA). Leaves were then 203 oven dried at 60°C for 48 hours and weighed to determine SLA. 204 Light availability 205 Light availability in the tropical forest understory is heterogeneous due to a mosaic of 206 canopy gaps and branch falls of differing ages and sizes. Because photosynthetic processes in 207 the understory are primarily light-limited (Pearcy 1988), we estimated light availability directly 208 above each of the 397 seedlings using hemispherical canopy photographs taken with a digital 209 camera (Coolpix 4500, Nikon Corp., Tokyo, Japan) mounted with a fisheye lens (Fisheye 210 Converter FC-E8 0.21x, Nikon Corp.). 211 Data analyses 212 We analyzed ETR light response curves for saturating photon flux density (PFDsat) using 213 Photosyn Assistant (version 1.1, Dundee Scientific, Dundee, UK) as described by Prioul and 214 Chartier (1977). Hemispheric canopy photographs were analyzed for total light transmittance 215 (Ttotal; proportion of above-canopy ambient) using Gap Light Analyzer (Frazer et al. 1999). We 216 used a general linear model in SAS (proc glm; version 9.2, SAS Institute Inc., Cary, North 217 Carolina, USA) to determine whether liana and tree seedlings overall were found in differing 218 Running head: Lianas always outperform trees (Pasquini et al.) 11 light environments. Physiological and morphological data were analyzed by mixed linear 219 models in SAS (proc mixed). The mixed linear model procedure was used rather than general 220 linear model procedure to yield Akaike Information Criterion (AIC) values for each model. 221 Models were run on individual leaf physiological and morphological measures with fixed main 222 effects of form (liana versus tree), species nested within form, single nutrient main effects (N, P, 223 K), two-way nutrient interactions (N × P, N × K, P × K), nutrient by growth form interactions (N 224 × form, P × form, K × form), and Ttotal (to control for heterogeneity of the light environment). 225 Random effects were statistical replicate (Rep) and block nested within replicate. For similar 226 analyses see Pasquini and Santiago (2012) and Wright et al. (2011). Models that included effects 227 of species as well as growth form were compared using AIC values (Appendix B: Tables B1-228 B6). Standard data transformations (natural log, square root, and arcsine) were performed to 229 meet the assumption of normality as determined by the Shapiro-Wilk W-statistic. Ttotal are 230 proportional data and were logit transformed accordingly (Warton and Hui 2011). To control for 231 type I error (α-error) in the multiple comparisons we used Family Discovery Rate (FDR) 232 corrections described by Pike (2011) to adjust significant P-values; FDR corrected values are 233 reported in the notes associated with each table. 234 235 Results 236 Model selection 237 Mixed linear models were performed in two different ways, with and without species 238 included. For all physiological and morphological variables, the model including species was a 239 better fit to the data as determined by comparing AIC values (Appendix B: Tables B1-B6). In 240 the results below, the findings based on the model with species included are presented. 241 Running head: Lianas always outperform trees (Pasquini et al.) 12 Physiological indices of performance 242 Lianas performed substantially better than trees for all physiological metrics (14-21% 243 greater; Table 1, Fig. 2A-C). As expected, seedling photosynthetic performance was affected by 244 light availability (Ttotal; Table 1) and thus it is important to note that mean understory light 245 availability did not differ between liana and tree seedlings (Lianas: 6.0 ± 0.6% SE, Trees: 5.8 ± 246 0.7% SE, F1,395 = 0.08, P = 0.78; Appendix C: Fig. C1). In addition, species within growth 247 forms differed significantly in physiological performance (Table 1). 248 Nutrient additions, especially P alone enhanced photosynthetic physiology, whereas N or 249 K addition never did. P addition increased ETRmax by 9.6% (Table 1, Fig. 3). P addition also 250 caused a marginally significant increase in qP (8.8%, P = 0.033; Table 1). Surprisingly, when K 251 was added in combination with P it decreased the benefit to performance caused by adding P 252 alone as indicated by consistent significant P × K interactions (Table 1, Fig. 4A-C). 253 Specifically, P and K together decreased the benefit of adding P alone for ETRmax, PFDsat, and qP 254 by 7.6%, 9.2%, and 10.2%, respectively (Table 1, Fig. 4A-C). Nutrient additions enhanced the 255 physiological performance of both lianas and trees to a similar degree (i.e., no significant 256 interaction between growth form and nutrient addition). For the effects of all nutrient treatment 257 combinations on the physiological responses of trees versus lianas see Appendix C (Fig. C2). 258 Plant architectural traits 259 Lianas and tree seedlings were significantly different from each other for all but one 260 metric of plant architecture (Table 2, Fig. 5A-F). Liana crowns were 32.0% deeper, their leaves 261 were 10.5% thicker, their internodes were 27.3% longer, and their petioles were 111.2% longer 262 than trees (Table 2, Fig. 5A-D). Tree leaf angles were 3.9% greater, and they had 9.2% greater 263 SLA than lianas (Table 2, Fig. 5E-F). Tree crowns were only marginally larger than liana 264 Running head: Lianas always outperform trees (Pasquini et al.) 13 crowns (12.9%, P = 0.024; Table 2). Surprisingly, light availability (Ttotal) had little impact on 265 seedling architecture except for SLA (Table 2). Species within growth forms differed 266 significantly in seedling architecture (Table 2). Liana seedlings averaged 28.2 ± 0.9 cm in height 267 and tree seedlings averaged 29.6 ± 1.0 cm in height (overall seedling height was 28.9 ± 0.7 cm). 268 Adding nutrients alone or in combination caused very few significant changes in seedling 269 morphology (Table 2, Fig. 6A-B). Specifically adding K caused a significant but small increase 270 (6.5%) in SLA, and P alone and K alone caused marginally significant but fairly substantial 271 increases in leaf angle (P: 10.6%, P = 0.046, K: 10.0%, P = 0.042; Table 2). If these results were 272 additive for P and K, then adding P and K together should have caused an even greater increase 273 in leaf angle; however this did not occur. Instead, leaf angles were close to control levels when P 274 and K were added together (significant P × K interaction; Table 2, Fig. 6B). We did detect one 275 case where nutrient additions caused the opposite response between lianas versus trees; P 276 addition caused petiole length to increase (15.6%) for lianas but decrease (15.7%) for trees 277 (significant growth form × P interaction; Table 2, Fig. 7). Nonetheless, the strong signal here is 278 that nutrient amendments had little impact on seven different metrics of seedling morphology. 279 For the effects of all nutrient treatment combinations on the architectural traits of trees versus 280 lianas see Appendix C (Figs. C3 and C4). 281 282 Discussion 283 To our knowledge, this is the first study to demonstrate that the early seedling stages of 284 common species of lianas substantially outperform (from 14-21%) common species of trees for 285 three key photosynthetic metrics regardless of macronutrient availability. Increasing nutrient 286 supply rates for P alone increased the performance of both lianas and tree seedlings to a similar 287 Running head: Lianas always outperform trees (Pasquini et al.) 14 degree but adding K with P dragged this performance benefit down. Regardless, the take home 288 message here is that long-term nutrient amendments never benefited lianas more than trees for 289 any macronutrient or any macronutrient combination. Also, and somewhat surprisingly, nitrogen 290 addition never caused any significant change in any physiological or morphological metric. In 291 addition, we were surprised that lianas and trees were architecturally quite different from each 292 other even during the free-standing seedling stages when they appear morphologically quite 293 similar (Putz 1983; Table 2, Fig. 5A-F). Nutrient enrichment did not change this in any way. 294 Thus these early morphological differences were robust and did not change even under long-term 295 and sharply contrasting soil nutrient supply rates. We suggest that our findings are broadly 296 applicable because we studied a phylogenetically diverse array of 13 species from 13 families. 297 Overall, our findings demonstrate that liana seedlings growing in deep shade are always capable 298 of higher photosynthetic performance than tree seedlings under ambient light levels and under 299 sharply contrasting levels of macronutrients (e.g., N vs. P vs. K) or under ambient nutrient levels. 300 Thus the advantage of having a liana growth habit occurs very early ontogenetically prior to any 301 use of a trellis for support. Our results provide strong evidence that P limits photosynthetic 302 performance of seedlings of both trees and lianas in deeply shaded understory habitats. 303 P limits photosynthetic performance but P and K together do not 304 Adding P caused a significant increase in one of three measures of photosynthetic 305 performance (ETRmax; Table 1, Fig. 3 and 4A) and a marginally significant increase in a second 306 measure (qp; Table 1, Fig. 4B). Adding K also increased photosynthetic performance, but this 307 increase was never significant (Table 1, Fig. 4A-C). Surprisingly adding P and K together 308 decreased photosynthetic performance relative to the addition of P alone (significant P × K 309 interaction) when it should have caused an increase in performance if the effect of each 310 Running head: Lianas always outperform trees (Pasquini et al.) 15 macronutrient alone was additive (Table 1, Fig. 4A-C). While the mechanistic basis of this is not 311 clear, we suggest that it is likely linked to alterations in stomatal control that occur with additions 312 of K. 313 Our results build on past studies that demonstrated that soil resources limit plant 314 performance even in deeply shaded habitats (Cai et al. 2008, Kaspari et al. 2008) but here we 315 identify which macronutrients were limiting or co-limiting. P addition enhanced ETRmax because 316 P is known to increase biochemical efficiency of the light reactions of photosynthesis and 317 promote enhanced carbon assimilation rates (Kirschbaum and Tompkins 1990, Raaimakers et al. 318 1995). Previous studies at this site demonstrate unequivocally that multiple soil resources co-319 limit trees in deep shade and we extend these results to seedlings of lianas. Thus, even a growth-320 form that is quite light-demanding and fast-growing can still be limited by soil resources when 321 light is at very low levels. We could not detect any impact of N additions on physiological 322 performance. Nonetheless, N, P, K, P × K, N × P and N × K all have been shown at times to 323 limit physiological performance, growth rate, or both, among woody species (current study, 324 Wright et al. 2011, Pasquini and Santiago 2012, Santiago et al. 2012). Still, our results strongly 325 point to P as the key limiting or co-limiting soil resource within the understory. Overall, we 326 demonstrate that light can no longer be considered the only limiting resource in deeply shaded 327 tropical habitats. 328 Liana and tree seedling architecture are markedly different 329 Lianas are classified as a separate growth form from trees because they are structural 330 parasites and require trellises to reach the canopy. Nonetheless, it was unknown whether key 331 architectural traits contrast between seedlings of lianas and trees prior to lianas acquiring a trellis 332 and prior to sending up searcher shoots. Here, counter to conventional wisdom, we show that 333 Running head: Lianas always outperform trees (Pasquini et al.) 16 liana seedlings differ for a suite of architectural traits. Nutrient additions rarely caused changes 334 in any of these traits at least while these seedlings were in deep shade. Thus differences in liana 335 architecture are expressed before lianas have located a trellis or before they rapidly increase 336 growth rates under conditions of higher light availability (Den Dubbelden and Oosterbeek 1995). 337 The height at which lianas begin to utilize external support is usually between 30-40 cm in 338 lowland tropical forests of Southeast Asia and Latin America (Putz and Holbrook 2009). The 339 average height of seedlings used in this study (29.1 cm) was close to this range, but all study 340 individuals were self-supporting. 341 We found that lianas had significantly lower SLA and thicker leaves compared to trees 342 (Table 2, Fig. 5B and F); in contrast, lianas that have reached the canopy typically have 343 significantly higher SLA and thinner leaves than trees (Lambers and Poorter 1992, Cai et al. 344 2009, Zhu and Cao 2010, Asner and Martin 2012, Santiago et al. 2015). Thus ontogenetic trait 345 shifts appear to be occurring for these important leaf structural traits. Leaves with low SLA are 346 more costly to build and high SLA is a characteristic of fast growing plants (Lambers and 347 Poorter 1992, Baruch and Goldstein 1999). Low SLA is also associated with both reduced 348 susceptibility to herbivores (Poorter et al. 2009) and increase leaf lifespan (Wright, I. J. et al. 349 2004). Low SLA and thick leaves of liana seedlings may allow them to persist for long periods 350 in the understory until they can access the canopy. 351 The physiology of lianas and their increase in Neotropical forests 352 There is compelling evidence that lianas are increasing in Neotropical forests (e.g., 353 Schnitzer 2015 and citations therein). Here we demonstrate that a phylogenetically diverse 354 group of lianas had enhanced physiological performance compared to a phylogenetically diverse 355 group of tree species (Table 1, Fig. 2A-C). Thus our findings extend previous research that 356 Running head: Lianas always outperform trees (Pasquini et al.) 17 found greater performance (Amax) by canopy-level lianas (Zhu and Cao 2009, Han et al. 2010, 357 Zhu and Cao 2010, Asner and Martin 2012, Santiago et al. 2015) to very early seedling stages in 358 the understory. We also demonstrated greater qP in liana seedlings relative to tree seedlings 359 (Table 1, Fig. 2B), which suggests that lianas are able to maximize the amount of incoming solar 360 radiation utilized for photosynthesis. This would confer an advantage in photosynthetic carbon 361 assimilation to lianas in rapidly changing light conditions seen in the understory due to short, 362 high intensity sunflecks (Chazdon 1988, Pearcy 1988). Overall our results suggest that any 363 changes in fertility, whether natural or anthropogenic, will not disproportionally favor lianas 364 because lianas already outperform trees regardless of fertility and lianas and trees responded 365 similarly to nutrient additions (only one significant growth form by nutrient interaction, Tables 1 366 and 2). Thus our results do not support nutrient loading as a mechanism of increasing liana 367 abundance in the Neotropics. Rather, our finding that lianas always outperform trees under 368 sharply contrasting rates of resource supply of macronutrients or their combination will allow 369 lianas to increase in abundance if disturbance rates are increasing in Neotropical forests as some 370 have suggested (Phillips et al. 2004). Moreover, an increase in liana abundance will likely lead 371 to lower forest-wide storage of carbon because lianas often displace trees and only replace 24% 372 of the biomass (Schnitzer et al. 2014). 373 374 Acknowledgements 375 We thank Omar Hernández, Rufino González and David Brassfield (Smithsonian Tropical 376 Research Institute) for help with plant identification and Juan Carrion (Universidad de Panamá), 377 and Eric Griffin (University of Pittsburgh) for field assistance. We would also like to thank 378 Steve Hovick (Ohio State University) and Jonathan Pruitt (University of Pittsburgh) for 379 Running head: Lianas always outperform trees (Pasquini et al.) 18 statistical advice, and Darrel Jenerette (University of California, Riverside), George Vourlitis 380 (Cal State San Marcos) and Walter Carson (University of Pittsburgh) for comments on earlier 381 drafts. We also thank two anonymous reviewers for their helpful comments. Funding for this 382 research was provided by the Department of Botany and Plant Sciences at the University of 383 California, Riverside and a STRI Short-Term Fellowship to SCP and by the Smithsonian 384 Scholarly Studies program to SJW. 385 Running head: Lianas always outperform trees (Pasquini et al.) 19 Literature Cited 386 Asner, G. P., and R. E. Martin. 2012. Contrasting leaf chemical traits in tropical lianas and trees: 387 implications for future forest composition. Ecology Letters 15:1001-1007. 388 Balfour, D. A., and W. J. Bond. 1993. Factors limiting climber distribution and abundance in a 389 southern African forest. Journal of Ecology 81:93-100. 390 Baruch, Z., and G. Goldstein. 1999. Leaf construction cost, nutrient concentration, and net CO2 391 assimilation of native and invasive species in Hawaii. Oecologia 121:183-192. 392 Benítez-Malvido, J., and M. Martínez-Ramos. 2003. Impact of forest fragmentation on 393 understory plant species richness in Amazonia. Conservation Biology 17:389-400. 394 Björkman, O., and B. Demmig. 1987. Photon yield of O2 evolution and chlorophyll fluorescence 395 characteristics at 77 K among vascular plants of diverse origins. Planta 170:489-504. 396 Bloom, A. J., F. S. Chapin, III, and H. A. Mooney. 1985. Resource limitation in plants – an 397 economic analogy. Annual Review of Ecology and Systematics 16:363-392. 398 Bruijnzeel, L. A., and J. Proctor. 1995. Hydrology and biogeochemistry of tropical montane 399 cloud forests: what do we really know? Pages 38-78 in L. S. Hamilton, J. O. Juvik, and F. 400 N. Scatena, editors. Tropical montane cloud forests. Springer-Verlag, New York. 401 Bunker, D. E., F. DeClerck, J. C. Bradford, R. K. Colwell, I. Perfecto, O. L. Phillips, M. 402 Sankaran, and S. Naeem. 2005. Species loss and aboveground carbon storage in a tropical 403 forest. Science 310:1029-1031. 404 Cai, Z.-Q., and F. Bongers. 2007. Contrasting nitrogen and phosphorus resorption efficiencies in 405 trees and lianas from a tropical montane rain forest in Xishuangbanna, south-west China. 406 Journal of Tropical Ecology 23:115-118. 407 Running head: Lianas always outperform trees (Pasquini et al.) 20 Cai, Z.-Q., L. Poorter, Q. Han, and F. Bongers. 2008. Effects of light and nutrients on seedlings 408 of tropical Bauhinia lianas and trees. Tree Physiology 28:1277-1285. 409 Cai, Z.-Q., S. A. Schnitzer, and F. Bongers. 2009. Seasonal differences in leaf-level physiology 410 give lianas a competitive advantage over trees in a tropical seasonal forest. Oecologia 411 161:25-33. 412 Castellanos, A. E., H. A. Mooney, S. H. Bullock, C. Jones, and R. Robichaux. 1989. Leaf, stem, 413 and metamer characteristics of vines in a tropical deciduous forest in Jalisco, Mexico. 414 Biotropica 21:41-49. 415 Ceccon, E., S. Sánchez, and J. Campo. 2004. Tree seedling dynamics in two abandoned tropical 416 dry forests of differing successional status in Yucatán, Mexico: a field experiment with N 417 and P fertilization. Plant Ecology 170:277-285. 418 Chave, J., J. Olivier, F. Bongers, P. Châtelet, P.-M. Forget, P. van der Meer, N. Norden, B. 419 Riéra, and P. Charles-Dominique. 2008. Above-ground biomass and productivity in a 420 rain forest of eastern South America. Journal of Tropical Ecology 24:355-366. 421 Chazdon, R. L. 1988. Sunflecks and their importance to forest understory plants. Advances in 422 Ecological Research 18:1-63. 423 Corre, M. D., E. Veldkamp, J. Arnold, and S. J. Wright. 2010. Impact of elevated N input on soil 424 N cycling and losses in old-growth lowland and montane forests in Panama. Ecology 425 91:1715-1729. 426 Den Dubbelden, K. C., and B. Oosterbeek. 1995. The availability of external support affects 427 allocation patterns and morphology of herbaceous climbing plants. Functional Ecology 428 9:628-634. 429 Running head: Lianas always outperform trees (Pasquini et al.) 21 Denslow, J. S., P. Vitousek, and J. C. Schultz. 1987. Bioassays of nutrient limitation in a tropical 430 rain forest soil. Oecologia 74:370-376. 431 Foster, J. R., P. A. Townsend, and C. E. Zganjar. 2008. Spatial and temporal patterns of gap 432 dominance by low-canopy lianas detected using EO-1 Hyperion and Landsat Thematic 433 Mapper. Remote Sensing of Environment 112:2104-2117. 434 Frazer, G. W., C. D. Canham, and K. P. Lertzman. 1999. Gap light analyzer (GLA), version 2.0: 435 imaging software to extract canopy structure and gap light indices from true-colour 436 fisheye photographs. Simon Fraser University, Burnaby, British Columbia, Canada, and 437 the Institute of Ecosystem Studies, Millbrook, New York, USA. 438 Garwood, N. C. 2009. Seedlings of Barro Colorado Island and the Neotropics. Cornell 439 University Press, Ithaca, New York, USA. 440 Gentry, A. H. 2009. The distribution and evolution of climbing plants. Pages 3-51 in F. E. Putz 441 and H. A. Mooney, editors. The biology of vines. Cambridge University Press, 442 Cambridge, UK. 443 Gerwing, J. J., and D. L. Farias. 2000. Integrating liana abundance and forest stature into an 444 estimate of total aboveground biomass for an eastern Amazonian forest. Journal of 445 Tropical Ecology 16:327-335. 446 Gilbert, B., S. J. Wright, H. C. Muller-Landau, K. Kitajima, and A. Hernandez. 2006. Life 447 history trade-offs in tropical trees and lianas. Ecology 87:1281-1288. 448 Han, L., L. J. Xie, K. J. Dai, Q. Yang, and Z. Q. Cai. 2010. Contrasting leaf characteristics of 449 trees and lianas in secondary and mature forests in southwestern China. Photosynthetica 450 48:559-566. 451 Running head: Lianas always outperform trees (Pasquini et al.) 22 Hättenschwiler, S. 2002. Liana seedling growth in response to fertilisation in a neotropical forest 452 understorey. Basic and Applied Ecology 3:135-143. 453 Hedin, L. O., E. N. J. Brookshire, D. N. L. Menge, and A. Barron. 2009. The nitrogen paradox in 454 tropical forest ecosystems. Annual Review of Ecology, Evolution and Systematics 455 40:613-635. 456 Hietz, P., B. L. Turner, W. Wanek, A. Richter, C. A. Nock, and S. J. Wright. 2011. Long-term 457 change in the nitrogen cycle of tropical forests. Science 334:664-666. 458 Holste, E. K., R. K. Kobe, and C. F. Vriesendorp. 2011. Seedling growth responses to soil 459 resources in the understory of a wet tropical forest. Ecology 92:1828-1838. 460 Kaspari, M., M. N. Garcia, K. E. Harms, M. Santana, S. J. Wright, and J. B. Yavitt. 2008. 461 Multiple nutrients limit litterfall and decomposition in a tropical forest. Ecology Letters 462 11:35-43. 463 Kirschbaum, M. U. F., and D. Tompkins. 1990. Photosynthetic responses to phosphorus nutrition 464 in Eucalyptus grandis seedlings. Australian Journal of Plant Physiology 17:527-535. 465 Kusumoto, B., and T. Enoki. 2008. Contribution of a liana species, Mucuna macrocarpa Wall., 466 to litterfall production and nitrogen input in a subtropical evergreen broad-leaved forest. 467 Journal of Forest Research 13:35-42. 468 Lambers, H., and H. Poorter. 1992. Inherent variation in growth rate between higher plants: a 469 search for physiological causes and ecological consequences. Advances in Ecological 470 Research 23:187-261. 471 Laurance, W. F. 2008. Environmental promise and peril in the Amazon. Pages 458-473 in W. P. 472 Carson and S. A. Schnitzer, editors. Tropical forest community ecology. Wiley-473 Blackwell, Oxford, UK. 474 Running head: Lianas always outperform trees (Pasquini et al.) 23 Maxwell, K., and G. N. Johnson. 2000. Chlorophyll fluorescence – a practical guide. Journal of 475 Experimental Botany 51:659-668. 476 Niklas, K. J. 1994. Comparisons among biomass allocation and spatial distribution patterns of 477 some vine, pteridophyte, and gymnosperm shoots. American Journal of Botany 81:1416-478 1421. 479 Pasquini, S. C., and L. S. Santiago. 2012. Nutrients limit photosynthesis in seedlings of a 480 lowland tropical forest tree species. Oecologia 168:311-319. 481 Pearcy, R. W. 1988. Photosynthetic utilisation of lightflecks by understory plants. Australian 482 Journal of Plant Physiology 15:223-238. 483 Pérez-Salicrup, D. R., V. L. Sork, and F. E. Putz. 2001. Lianas and trees in a liana forest of 484 Amazonian Bolivia. Biotropica 33:34-47. 485 Phillips, O. L., T. R. Baker, L. Arroyo, N. Higuchi, T. J. Killeen, W. F. Laurance, S. L. Lewis, J. 486 Lloyd, Y. Malhi, A. Monteagudo, D. A. Neill, P. Núñez Vargas, J. N. M. Silva, J. 487 Terborgh, R. Vásquez Martínez, M. Alexiades, S. Almeida, S. Brown, J. Chave, J. A. 488 Comiskey, C. I. Czimczik, A. Di Fiore, T. Erwin, C. Kuebler, S. G. Laurance, H. E. M. 489 Nascimento, J. Olivier, W. Palacios, S. Patino, N. C. A. Pitman, C. A. Quesada, M. 490 Saldias, A. Torres Lezama, and B. Vinceti. 2004. Pattern and process in Amazon tree 491 turnover, 1976-2001. Philosophical Transactions of the Royal Society B 359:381-407. 492 Phillips, O. L., R. V. Martínez, L. Arroyo, T. B. Baker, T. Killeen, S. L. Lewis, Y. Malhi, A. M. 493 Mendoza, D. Neill, P. N. Vargas, M. Alexiades, C. Cerón, A. Di Fiore, T. Erwin, A. 494 Jardim, W. Palacios, M. Saldias, and B. Vinceti. 2002. Increasing dominance of large 495 lianas in Amazonian forests. Nature 418:770-774. 496 Running head: Lianas always outperform trees (Pasquini et al.) 24 Pike, N. 2011. Using False Discovery Rates for multiple comparisons in ecology and evolution. 497 Methods in Ecology and Evolution 2:278-282. 498 Poorter, H., Ü. Niinemets, L. Poorter, I. J. Wright, and R. Villar. 2009. Causes and consequences 499 of variation in leaf mass per area (LMA): a meta-analysis. New Phytologist 182:565-588. 500 Prioul, J. L., and P. Chartier. 1977. Partitioning of transfer and carboxylation components of 501 intracellular resistance to photosynthetic CO2 fixation: a critical analysis of the methods 502 used. Annals of Botany 41:789-800. 503 Proctor, J., J. M. Anderson, P. Chai, and H. W. Vallack. 1983. Ecological studies in four 504 contrasting lowland rain forests in Gunung Mulu National Park, Sarawak. I. Forest 505 environment, structure and floristics. Journal of Ecology 71:237-260. 506 Putz, F. E. 1983. Liana biomass and leaf area of a "tierra firme" forest in the Rio Negro Basin, 507 Venezuela. Biotropica 15:185-189. 508 Putz, F. E. 1984. The natural history of lianas on Barro Colorado Island, Panama. Ecology 509 65:1713-1724. 510 Putz, F. E. 1985. Woody vines and forest management in Malaysia. Commonwealth Forestry 511 Review 64:1713-1724. 512 Putz, F. E., and P. Chai. 1987. Ecological studies of lianas in Lambir National Park, Sarawak, 513 Malaysia. Journal of Ecology 75:523-531. 514 Putz, F. E., and N. M. Holbrook. 2009. Biomechanical studies of vines. Pages 73-97 in F. E. Putz 515 and H. A. Mooney, editors. The biology of vines. Cambridge University Press, 516 Cambridge, UK. 517 Running head: Lianas always outperform trees (Pasquini et al.) 25 Raaimakers, D., R. G. A. Boot, P. Dijkstra, S. Pot, and T. Pons. 1995. Photosynthetic rates in 518 relation to leaf phosphorus content in pioneer versus climax tropical rainforest trees. 519 Oecologia 102:120-125. 520 Santiago, L. S., S. C. Pasquini, and M. E. De Guzman. 2015. Physiological implications of the 521 liana growth form. Pages 288-298 in S. A. Schnitzer, F. Bongers, R. J. Burnham, and F. 522 E. Putz, editors. The ecology of lianas. Wiley-Blackwell Publishing, Oxford, UK. 523 Santiago, L. S., and S. J. Wright. 2007. Leaf functional traits of tropical forest plants in relation 524 to growth form. Functional Ecology 21:19-27. 525 Santiago, L. S., S. J. Wright, K. E. Harms, J. B. Yavitt, C. Korine, M. N. Garcia, and B. L. 526 Turner. 2012. Tropical tree seedling growth responses to nitrogen, phosphorus and 527 potassium addition. Journal of Ecology 100:309-316. 528 Schnitzer, S. A. 2015. Increasing liana abundance in Neotropical forests: causes and 529 consequences. Pages 451-464 in S. A. Schnitzer, F. Bongers, R. J. Burnham, and F. E. 530 Putz, editors. The ecology of lianas. Wiley-Blackwell, Oxford, UK. 531 Schnitzer, S. A., and F. Bongers. 2011. Increasing liana abundance and biomass in tropical 532 forests: emerging patterns and putative mechanisms. Ecology Letters 14:397-406. 533 Schnitzer, S. A., and W. P. Carson. 2010. Lianas suppress tree regeneration and diversity in 534 treefall gaps. Ecology Letters 13:849-857. 535 Schnitzer, S. A., J. W. Dalling, and W. P. Carson. 2000. The impact of lianas on tree 536 regeneration in tropical forest canopy gaps: evidence for an alternative pathway of gap-537 phase regeneration. Journal of Ecology 88:655-666. 538 Schnitzer, S. A., S. A. Mangan, J. W. Dalling, C. A. Baldeck, S. P. Hubbell, A. Ledo, H. C. 539 Muller-Landau, M. F. Tobin, S. Aguilar, D. Brassfield, A. Hernandez, S. Lao, R. Perez, 540 Running head: Lianas always outperform trees (Pasquini et al.) 26 O. Valdes, and S. R. Yorke. 2012. Liana abundance, diversity, and distribution on Barro 541 Colorado Island, Panama. PLoS One 7:e52114. 542 Schnitzer, S. A., G. van der Heijden, J. Mascaro, and W. P. Carson. 2014. Lianas in gaps reduce 543 carbon accumulation in a tropical forest. Ecology 95:3008-3017. 544 Stemke, J. A., and L. S. Santiago. 2011. Consequences of light absorptance in calculating 545 electron transport rate of desert and succulent plants. Photosynthetica 49:195-200. 546 Suzuki, W. 1987. Comparative ecology of Rubus species (Roseaceae). I. Ecological distribution 547 and life history characteristics of three species, R. palmatus var. coptophyllus, R. 548 microhyllus and R. crataegifolius. Plant Species Biology 2:85-100. 549 Turner, B. L., J. B. Yavitt, K. E. Harms, M. N. Garcia, T. E. Romero, and S. J. Wright. 2012. 550 Seasonal changes and treatment effects on soil inorganic nutrients following a decade of 551 fertilizer addition in a lowland tropical forest. Soil Science Society of America Journal 552 77:1357-1369. 553 Warton, D. I., and F. K. C. Hui. 2011. The arcsine is asinine: the analysis of proportions in 554 ecology. Ecology 92:3-10. 555 Winer, B. J., D. R. Brown, and K. M. Michels. 1991. Statistical principles in experimental 556 design. 3rd edition. McGraw-Hill, New York, New York, USA. 557 Wright, I. J., P. B. Reich, M. Westoby, D. D. Ackerly, Z. Baruch, F. Bongers, J. Cavender-Bares, 558 F. S. Chapin, III, J. H. C. Cornelissen, M. Diemer, J. Flexas, E. Garnier, P. K. Groom, J. 559 Gulias, K. Hikosaka, B. B. Lamont, T. Lee, W. Lee, C. Lusk, J. J. Midgley, M.-L. Navas, 560 Ü. Niinemets, J. Oleksyn, N. Osada, H. Poorter, P. Poot, L. Prior, V. Pyankov, C. 561 Roumet, S. C. Thomas, M. G. Tjoelker, E. J. Veneklaas, and R. Villar. 2004. The 562 worldwide leaf economics spectrum. Nature 428:821-827. 563 Running head: Lianas always outperform trees (Pasquini et al.) 27 Wright, S. J. 2010. The future of tropical forests. Annals of the New York Academy of Sciences 564 1195:1-27. 565 Wright, S. J., O. Calderdón, A. Hernandéz, and S. Paton. 2004. Are lianas increasing in 566 importance in tropical forests? A 17-year record from Panama. Ecology 85:484-489. 567 Wright, S. J., J. B. Yavitt, N. Wurzburger, B. L. Turner, E. V. J. Tanner, E. J. Sayer, L. S. 568 Santiago, M. Kaspari, L. O. Hedin, K. E. Harms, M. N. Garcia, and M. D. Corre. 2011. 569 Potassium, phosphorus, or nitrogen limit root allocation, tree growth, or litter production 570 in a lowland tropical forest. Ecology 92:1616-1625. 571 Yavitt, J. B., K. E. Harms, M. N. Garcia, S. J. Wright, F. He, and M. J. Mirabello. 2009. Spatial 572 heterogeneity of soil chemical properties in a lowland tropical moist forest, Panama. 573 Australian Journal of Soil Research 47:674-687. 574 Yorke, S. R., S. A. Schnitzer, J. Mascaro, S. G. Letcher, and W. P. Carson. 2013. Increasing 575 liana abundance and basal area in a tropical forest: the contribution of long-distance 576 clonal colonization. Biotropica 45:317-324. 577 Zhu, S.-D., and K.-F. Cao. 2009. Hydraulic properties and photosynthetic rates in co-occurring 578 lianas and trees in a seasonal tropical rainforest in southwestern China. Plant Ecology 579 204:295-304. 580 Zhu, S.-D., and K.-F. Cao. 2010. Contrasting cost-benefit strategy between lianas and trees in a 581 tropical seasonal rainforest in southwestern China. Oecologia 163:591-599. 582 583 Running head: Lianas always outperform trees (Pasquini et al.) 28 APPENDIX A 584 Figure showing map of study area. 585 586 APPENDIX B 587 Mixed linear model results for the factorial, incomplete-block design comparing statistical 588 model that includes species and statistical model that does not include species. 589 590 APPENDIX C 591 Figures showing the average of total transmission of light above liana and tree seedlings 592 used in study, and responses to full-factorial nitrogen, phosphorus, and potassium addition 593 compared to unfertilized controls for physiological and morphological traits. 594 Running head: Lianas always outperform trees (Pasquini et al.) 29 Table 1. Mixed linear model results of nitrogen (N), phosphorus (P), and potassium (K) 595 fertilization effects on leaf physiological traits for liana and tree seedlings. 596 ETRmax qP PFDsat Form 0.0003 <0.0001 0.0006 Species (Form) <0.0001 <0.0001 <0.0001 Ttotal 0.0012 <0.0001 0.0014 N 0.8286 0.3554 0.9483 P 0.0129 0.0333 0.3415 K 0.5808 0.1896 0.4405 N × P 0.0552 0.2328 0.1382 N × K 0.6880 0.9682 0.6790 P × K 0.0079 0.0112 0.0126 N × Form 0.1400 0.3081 0.4305 P × Form 0.6397 0.1915 0.7753 K × Form 0.9657 0.6398 0.9068 Sample size 394 393 394 Data presented are P-values for fixed effects. Bolded values are statistically significant using the 597 Family Discovery Rate (FDR) corrected P-value (P < 0.0237). Total light transmission (Ttotal; 598 proportion of above-canopy ambient), maximum electron transport rate (ETRmax; μmol m-2 s-1), 599 photochemical quenching (qP; unitless), and saturating photon flux density (PFDsat; μmol m-2 s-600 1). 601 R un ni ng h ea d: L ia na s al w ay s ou tp er fo rm tr ee s (P as qu in i e t a l.) 30 T ab le 2 . M ix ed li ne ar m od el re su lts o f n itr og en (N ), ph os ph or us (P ), an d po ta ss iu m (K ) f er til iz at io n ef fe ct s on m or ph ol og ic al tr ai ts 60 2 fo r l ia na a nd tr ee s ee dl in gs . 60 3 R el . d ep th R el . a re a T hi ck ne ss A ng le In te rn od e Pe tio le SL A Fo rm < 0. 00 01 0. 02 39 < 0. 00 01 0. 00 66 0. 00 28 < 0. 00 01 < 0. 00 01 Sp ec ie s (F or m ) < 0. 00 01 < 0. 00 01 < 0. 00 01 < 0. 00 01 < 0. 00 01 < 0. 00 01 < 0. 00 01 T t ot al 0. 90 91 0. 78 75 0. 67 00 0. 29 57 0. 68 46 0. 31 55 0. 00 99 N 0. 45 38 0. 35 45 0. 52 13 0. 27 38 0. 44 07 0. 20 33 0. 38 67 P 0. 08 50 0. 30 86 0. 26 69 0. 04 61 0. 40 54 0. 97 43 0. 64 09 K 0. 49 31 0. 59 22 0. 83 57 0. 04 20 0. 45 19 0. 25 00 0. 00 07 N × P 0. 69 73 0. 28 65 0. 71 66 0. 24 18 0. 02 95 0. 89 15 0. 18 89 N × K 0. 93 49 0. 51 07 0. 13 59 0. 34 54 0. 51 80 0. 07 93 0. 17 29 P × K 0. 85 40 0. 87 86 0. 26 92 0. 00 02 0. 88 28 0. 43 62 0. 16 63 N × F or m 0. 14 43 0. 06 73 0. 93 26 0. 76 39 0. 94 54 0. 84 91 0. 49 90 P × Fo rm 0. 66 41 0. 08 79 0. 68 36 0. 17 87 0. 58 60 0. 00 91 0. 09 75 K × F or m 0. 90 91 0. 19 72 0. 92 99 0. 97 15 0. 22 06 0. 09 01 0. 07 10 Sa m pl e si ze 38 8 39 0 38 1 38 2 38 3 39 0 39 3 D at a pr es en te d ar e P -v al ue s fo r f ix ed e ff ec ts . B ol de d va lu es a re s ta tis tic al ly s ig ni fi ca nt u si ng th e Fa m ily D is co ve ry R at e (F D R ) 60 4 co rr ec te d P -v al ue (P < 0 .0 11 5) . T ot al li gh t t ra ns m is si on (T to ta l; pr op or tio n of a bo ve -c an op y am bi en t) , r el at iv e cr ow n de pt h (c m ), 60 5 R un ni ng h ea d: L ia na s al w ay s ou tp er fo rm tr ee s (P as qu in i e t a l.) 31 re la tiv e cr ow n ar ea (c m 2 ) , l ea f t hi ck ne ss (m m ), le af a ng le (d eg re es fr om m ai n st em ), in te rn od e le ng th (m m ), pe tio le le ng th (m m ), an d 60 6 sp ec if ic le af a re a (S L A ; m 2 / kg ). 60 7 Running head: Lianas always outperform trees (Pasquini et al.) 32 Figure Legends 608 Fig. 1. Example chlorophyll fluorescence light response curve generated by the Mini-PAM 609 photosynthesis yield analyzer showing electron transport rate (ETR; solid line) and 610 photochemical quenching (qP; dashed line), for increasing levels of photon flux density (PFD). 611 612 Fig. 2. Significant main effects of growth form for (A) maximum electron transport rate 613 (ETRmax), (B) photochemical quenching coefficient (qP), and (C) saturating photon flux density 614 (PFDsat). Lianas and trees are represented by L and black bars, and T and open bars, 615 respectively. All nutrient treatments are pooled. Bars represent means (± 1 SE, N = 32 plots). 616 617 Fig. 3. Significant main effect of P addition for maximum electron transport rate (ETRmax). 618 Liana and tree seedlings are pooled. Treatments without P (C, N, K, and NK) and treatments 619 with P (P, NP, PK, and NPK) are also pooled. Bars represent means (± 1 SE, N = 16 plots). 620 621 Fig. 4. Significant P × K interactions for (A) maximum electron transport rate (ETRmax), (B) 622 photochemical quenching coefficient (qP), and (C) saturating photon flux density (PFDsat). Liana 623 and tree seedlings are pooled. Treatments without P or K (C and N), P treatments (P and NP), K 624 treatments (K and NK) and treatments with both P and K (PK and NPK) are also pooled. Bars 625 represent means (± 1 SE, N = 8 plots). 626 627 Fig. 5. Significant main effects of growth form for (A) relative crown depth, (B) leaf thickness, 628 (C) internode length, (D) petiole length, (E) leaf angle, and (F) specific leaf area (SLA). Lianas 629 and trees are represented by L and black bars, and T and open bars, respectively. Bars represent 630 Running head: Lianas always outperform trees (Pasquini et al.) 33 means (± 1 SE, N = 32 plots). 631 632 Fig. 6. Significant main effect of K for (A) specific leaf area (SLA) and significant interaction of 633 P × K for (B) leaf angle. Lianas and tree seedlings are pooled and represented by L and T, 634 respectively. Panel A represents a significant main effect of K, where treatments without K (C, 635 N, P, and NP) and treatments with K (K, NK, PK, and NPK) are pooled (N = 16 plots). Panel B 636 represents a significant P × K interaction where treatments without P or K (C and N), P 637 treatments (P and NP), K treatments (K and NK) and treatments with both P and K (PK and 638 NPK) are pooled (N = 8 plots). Bars represent means (± 1 SE). 639 640 Fig. 7. Significant interaction of P × growth form for petiole length. Lianas and trees are 641 represented by L and black bars, and T and open bars, respectively. Treatments without P (C, N, 642 K, and NK) and treatments with P (P and PK, NP and NPK) are pooled by growth form (N = 16 643 plots). Bars represent means (± 1 SE). 644 Running head: Lianas always outperform trees (Pasquini et al.) 34 Figure 1 645 646 Running head: Lianas always outperform trees (Pasquini et al.) 35 Figure 2 647 648 Running head: Lianas always outperform trees (Pasquini et al.) 36 Figure 3 649 650 651 Running head: Lianas always outperform trees (Pasquini et al.) 37 Figure 4 652 653 Running head: Lianas always outperform trees (Pasquini et al.) 38 Figure 5 654 655 Running head: Lianas always outperform trees (Pasquini et al.) 39 Figure 6 656 657 Running head: Lianas always outperform trees (Pasquini et al.) 40 Figure 7 658 659